Merge branch 'master' into hn-activation

Author: guaneec

localizations/fr-FR.json

@@ -7,6 +7,7 @@
 from fastapi import Body, APIRouter, HTTPException
 from fastapi.responses import JSONResponse
 from pydantic import BaseModel, Field, Json
+from typing import List
 import json
 import io
 import base64
@@ -15,12 +16,12 @@
 sampler_to_index = lambda name: next(filter(lambda row: name.lower() == row[1].name.lower(), enumerate(all_samplers)), None)
 
 class TextToImageResponse(BaseModel):
-    images: list[str] = Field(default=None, title="Image", description="The generated image in base64 format.")
+    images: List[str] = Field(default=None, title="Image", description="The generated image in base64 format.")
     parameters: Json
     info: Json
 
 class ImageToImageResponse(BaseModel):
-    images: list[str] = Field(default=None, title="Image", description="The generated image in base64 format.")
+    images: List[str] = Field(default=None, title="Image", description="The generated image in base64 format.")
     parameters: Json
     info: Json
 
@@ -65,7 +66,7 @@ def text2imgapi(self, txt2imgreq: StableDiffusionTxt2ImgProcessingAPI):
             i.save(buffer, format="png")
             b64images.append(base64.b64encode(buffer.getvalue()))
 
-        return TextToImageResponse(images=b64images, parameters=json.dumps(vars(txt2imgreq)), info=json.dumps(processed.info))
+        return TextToImageResponse(images=b64images, parameters=json.dumps(vars(txt2imgreq)), info=processed.js())
 
 
 
@@ -111,7 +112,11 @@ def img2imgapi(self, img2imgreq: StableDiffusionImg2ImgProcessingAPI):
             i.save(buffer, format="png")
             b64images.append(base64.b64encode(buffer.getvalue()))
 
-        return ImageToImageResponse(images=b64images, parameters=json.dumps(vars(img2imgreq)), info=json.dumps(processed.info))
+        if (not img2imgreq.include_init_images):
+            img2imgreq.init_images = None
+            img2imgreq.mask = None
+
+        return ImageToImageResponse(images=b64images, parameters=json.dumps(vars(img2imgreq)), info=processed.js())
 
     def extrasapi(self):
         raise NotImplementedError

modules/api/api.py

@@ -31,6 +31,7 @@ class ModelDef(BaseModel):
     field_alias: str
     field_type: Any
     field_value: Any
+    field_exclude: bool = False
 
 
 class PydanticModelGenerator:
@@ -78,15 +79,16 @@ def merge_class_params(class_):
                 field=underscore(fields["key"]), 
                 field_alias=fields["key"], 
                 field_type=fields["type"],
-                field_value=fields["default"]))
+                field_value=fields["default"],
+                field_exclude=fields["exclude"] if "exclude" in fields else False))
 
     def generate_model(self):
         """
         Creates a pydantic BaseModel
         from the json and overrides provided at initialization
         """
         fields = {
-            d.field: (d.field_type, Field(default=d.field_value, alias=d.field_alias)) for d in self._model_def
+            d.field: (d.field_type, Field(default=d.field_value, alias=d.field_alias, exclude=d.field_exclude)) for d in self._model_def
         }
         DynamicModel = create_model(self._model_name, **fields)
         DynamicModel.__config__.allow_population_by_field_name = True
@@ -102,5 +104,5 @@ def generate_model(self):
 StableDiffusionImg2ImgProcessingAPI = PydanticModelGenerator(
     "StableDiffusionProcessingImg2Img", 
     StableDiffusionProcessingImg2Img,
-    [{"key": "sampler_index", "type": str, "default": "Euler"}, {"key": "init_images", "type": list, "default": None}, {"key": "denoising_strength", "type": float, "default": 0.75}, {"key": "mask", "type": str, "default": None}]
+    [{"key": "sampler_index", "type": str, "default": "Euler"}, {"key": "init_images", "type": list, "default": None}, {"key": "denoising_strength", "type": float, "default": 0.75}, {"key": "mask", "type": str, "default": None}, {"key": "include_init_images", "type": bool, "default": False, "exclude" : True}]
 ).generate_model()

modules/api/models.py

@@ -5,6 +5,7 @@
 import os
 import sys
 import traceback
+import inspect
 
 import modules.textual_inversion.dataset
 import torch
@@ -15,20 +16,25 @@
 from modules.textual_inversion import textual_inversion
 from modules.textual_inversion.learn_schedule import LearnRateScheduler
 from torch import einsum
+from torch.nn.init import normal_, xavier_normal_, xavier_uniform_, kaiming_normal_, kaiming_uniform_, zeros_
 
 from collections import defaultdict, deque
 from statistics import stdev, mean
 
+
 class HypernetworkModule(torch.nn.Module):
     multiplier = 1.0
     activation_dict = {
         "relu": torch.nn.ReLU,
         "leakyrelu": torch.nn.LeakyReLU,
         "elu": torch.nn.ELU,
         "swish": torch.nn.Hardswish,
+        "tanh": torch.nn.Tanh,
+        "sigmoid": torch.nn.Sigmoid,
     }
+    activation_dict.update({cls_name.lower(): cls_obj for cls_name, cls_obj in inspect.getmembers(torch.nn.modules.activation) if inspect.isclass(cls_obj) and cls_obj.__module__ == 'torch.nn.modules.activation'})
 
-    def __init__(self, dim, state_dict=None, layer_structure=None, activation_func=None, add_layer_norm=False, use_dropout=False, activate_output=False):
+    def __init__(self, dim, state_dict=None, layer_structure=None, activation_func=None, weight_init='Normal', add_layer_norm=False, use_dropout=False, activate_output=False):
         super().__init__()
 
         assert layer_structure is not None, "layer_structure must not be None"
@@ -65,9 +71,24 @@ def __init__(self, dim, state_dict=None, layer_structure=None, activation_func=N
         else:
             for layer in self.linear:
                 if type(layer) == torch.nn.Linear or type(layer) == torch.nn.LayerNorm:
-                    layer.weight.data.normal_(mean=0.0, std=0.01)
-                    layer.bias.data.zero_()
-
+                    w, b = layer.weight.data, layer.bias.data
+                    if weight_init == "Normal" or type(layer) == torch.nn.LayerNorm:
+                        normal_(w, mean=0.0, std=0.01)
+                        normal_(b, mean=0.0, std=0.005)
+                    elif weight_init == 'XavierUniform':
+                        xavier_uniform_(w)
+                        zeros_(b)
+                    elif weight_init == 'XavierNormal':
+                        xavier_normal_(w)
+                        zeros_(b)
+                    elif weight_init == 'KaimingUniform':
+                        kaiming_uniform_(w, nonlinearity='leaky_relu' if 'leakyrelu' == activation_func else 'relu')
+                        zeros_(b)
+                    elif weight_init == 'KaimingNormal':
+                        kaiming_normal_(w, nonlinearity='leaky_relu' if 'leakyrelu' == activation_func else 'relu')
+                        zeros_(b)
+                    else:
+                        raise KeyError(f"Key {weight_init} is not defined as initialization!")
         self.to(devices.device)
 
     def fix_old_state_dict(self, state_dict):
@@ -105,7 +126,7 @@ class Hypernetwork:
     filename = None
     name = None
 
-    def __init__(self, name=None, enable_sizes=None, layer_structure=None, activation_func=None, add_layer_norm=False, use_dropout=False, activate_output=False):
+    def __init__(self, name=None, enable_sizes=None, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False, activate_output=False)
         self.filename = None
         self.name = name
         self.layers = {}
@@ -114,14 +135,15 @@ def __init__(self, name=None, enable_sizes=None, layer_structure=None, activatio
         self.sd_checkpoint_name = None
         self.layer_structure = layer_structure
         self.activation_func = activation_func
+        self.weight_init = weight_init
         self.add_layer_norm = add_layer_norm
         self.use_dropout = use_dropout
         self.activate_output = activate_output
 
         for size in enable_sizes or []:
             self.layers[size] = (
-                HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.add_layer_norm, self.use_dropout, self.activate_output),
-                HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.add_layer_norm, self.use_dropout, self.activate_output),
+                HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.weight_init, self.add_layer_norm, self.use_dropout, self.activate_output),
+                HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.weight_init, self.add_layer_norm, self.use_dropout, self.activate_output),
             )
 
     def weights(self):
@@ -145,6 +167,7 @@ def save(self, filename):
         state_dict['layer_structure'] = self.layer_structure
         state_dict['activation_func'] = self.activation_func
         state_dict['is_layer_norm'] = self.add_layer_norm
+        state_dict['weight_initialization'] = self.weight_init
         state_dict['use_dropout'] = self.use_dropout
         state_dict['sd_checkpoint'] = self.sd_checkpoint
         state_dict['sd_checkpoint_name'] = self.sd_checkpoint_name
@@ -160,16 +183,22 @@ def load(self, filename):
         state_dict = torch.load(filename, map_location='cpu')
 
         self.layer_structure = state_dict.get('layer_structure', [1, 2, 1])
+        print(self.layer_structure)
         self.activation_func = state_dict.get('activation_func', None)
+        print(f"Activation function is {self.activation_func}")
+        self.weight_init = state_dict.get('weight_initialization', 'Normal')
+        print(f"Weight initialization is {self.weight_init}")
         self.add_layer_norm = state_dict.get('is_layer_norm', False)
-        self.use_dropout = state_dict.get('use_dropout', False)
+        print(f"Layer norm is set to {self.add_layer_norm}")
+        self.use_dropout = state_dict.get('use_dropout', False
+        print(f"Dropout usage is set to {self.use_dropout}" )
         self.activate_output = state_dict.get('activate_output', True)
 
         for size, sd in state_dict.items():
             if type(size) == int:
                 self.layers[size] = (
-                    HypernetworkModule(size, sd[0], self.layer_structure, self.activation_func, self.add_layer_norm, self.use_dropout, self.activate_output),
-                    HypernetworkModule(size, sd[1], self.layer_structure, self.activation_func, self.add_layer_norm, self.use_dropout, self.activate_output),
+                    HypernetworkModule(size, sd[0], self.layer_structure, self.activation_func, self.weight_init, self.add_layer_norm, self.use_dropout, self.activate_output),
+                    HypernetworkModule(size, sd[1], self.layer_structure, self.activation_func, self.weight_init, self.add_layer_norm, self.use_dropout, self.activate_output),
                 )
 
         self.name = state_dict.get('name', self.name)

modules/hypernetworks/hypernetwork.py

@@ -8,8 +8,9 @@
 from modules import devices, sd_hijack, shared
 from modules.hypernetworks import hypernetwork
 
+keys = list(hypernetwork.HypernetworkModule.activation_dict.keys())
 
-def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, activation_func=None, add_layer_norm=False, use_dropout=False):
+def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False):
     # Remove illegal characters from name.
     name = "".join( x for x in name if (x.isalnum() or x in "._- "))
 
@@ -25,6 +26,7 @@ def create_hypernetwork(name, enable_sizes, overwrite_old, layer_structure=None,
         enable_sizes=[int(x) for x in enable_sizes],
         layer_structure=layer_structure,
         activation_func=activation_func,
+        weight_init=weight_init,
         add_layer_norm=add_layer_norm,
         use_dropout=use_dropout,
     )

modules/hypernetworks/ui.py

@@ -277,7 +277,7 @@ def resize(im, w, h):
 invalid_filename_prefix = ' '
 invalid_filename_postfix = ' .'
 re_nonletters = re.compile(r'[\s' + string.punctuation + ']+')
-re_pattern = re.compile(r"([^\[\]]+|\[([^]]+)]|[\[\]]*)")
+re_pattern = re.compile(r"(.*?)(?:\[([^\[\]]+)\]|$)")
 re_pattern_arg = re.compile(r"(.*)<([^>]*)>$")
 max_filename_part_length = 128
 
@@ -343,7 +343,7 @@ def prompt_words(self):
     def datetime(self, *args):
         time_datetime = datetime.datetime.now()
 
-        time_format = args[0] if len(args) > 0 else self.default_time_format
+        time_format = args[0] if len(args) > 0 and args[0] != "" else self.default_time_format
         try:
             time_zone = pytz.timezone(args[1]) if len(args) > 1 else None
         except pytz.exceptions.UnknownTimeZoneError as _:
@@ -362,9 +362,9 @@ def apply(self, x):
 
         for m in re_pattern.finditer(x):
             text, pattern = m.groups()
+            res += text
 
             if pattern is None:
-                res += text
                 continue
 
             pattern_args = []
@@ -385,12 +385,9 @@ def apply(self, x):
                     print(f"Error adding [{pattern}] to filename", file=sys.stderr)
                     print(traceback.format_exc(), file=sys.stderr)
 
-                if replacement is None:
-                    res += f'[{pattern}]'
-                else:
+                if replacement is not None:
                     res += str(replacement)
-
-                continue
+                    continue
 
             res += f'[{pattern}]'